The incidence of restenosis after a first successful percutaneous transluminal angioplasty of a native renal artery and the clinical and angiographic variables that may influence its occurrence were studied in 104 hypertensive patients. Angiograms obtained immediately before and after angioplasty and, in 92 patients, 8.8 +/- 6.0 months after angioplasty were interpreted separately by two observers. Stenosis severity was classified into five grades, and restenosis was defined by a stenosis one grade or more higher at follow-up than immediately after angioplasty. Interobserver concordance for etiology, stenosis grade, and other angiographic items yielded kappa coefficients in the range of 0.328-0.942. Sessions were organized to reach a consensus in each case. Ostial stenoses were more frequent in patients with atheromatous stenoses, and branch stenoses were more frequent in those with fibromuscular dysplasia. There was no significant difference between the 15 patients (16%) with restenosis and those without concerning sex distribution, mean age, mean blood pressure, plasma creatinine level, and etiology distribution. Truncal stenoses were less prone to restenosis than ostial or branch stenoses (12% versus 35%, respectively; 95% confidence interval of difference, -0.6% to 47%). In patients with atheromatous stenoses, aortitis or aortic ectasia were associated with a high restenosis incidence (35% when present versus 8% when absent; 95% confidence interval of difference, 5% to 48%). In conclusion, restenosis was observed in one sixth of patients after a first successful renal angioplasty; its incidence was low in patients with truncal stenoses and high in those with severe aortic atheroma. Automated renal artery stenosis quantification methods are needed to standardize stenosis description.